grok-imagine-api
Pass
Audited by Gen Agent Trust Hub on Sep 21, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [SAFE]: The skill provides integration guidance and code snippets for using the xAI Grok Imagine API across multiple integration paths including official SDKs and REST endpoints.
- [SAFE]: Credential management follows best practices by using environment variable placeholders and documentation strings for API keys.
- [SAFE]: Network and file system operations are limited to the intended functionality of interacting with the xAI API and processing image files locally.
- [INDIRECT_PROMPT_INJECTION]: The skill processes external data (prompts and image URLs) which are sent to an external model provider. 1. Ingestion points: User-provided prompt and image_url parameters in SKILL.md and references/image-api.md. 2. Boundary markers: None identified; user input is interpolated directly into API request structures. 3. Capability inventory: Includes network access to api.x.ai and local file system write access for saving generated images, as seen in SKILL.md and references/image-api.md. 4. Sanitization: Not demonstrated in the provided code examples.
Audit Metadata